Troponin I and T similarly predicted in-hospital ischemic events (AUC ~0.63), TnT better predicted in-hospital mortality (AUC 0.63 vs 0.54, p=0.03), GRACE was superior overall.
Does Troponin I or Troponin T offer better prediction of in-hospital ischemic events, bleeding events, and mortality in patients with acute coronary syndrome?
Troponin I and T have similar, modest predictive value for ischemic and bleeding events in ACS, while TnT is slightly better for in-hospital mortality, though clinical scores like GRACE remain superior.
Absolute Event Rate: 0% vs 0%
Abstract Introduction Most medical facilities use high-sensitivity troponin for assessing patients presenting with chest pain. However, variations exist among assays and manufacturers (I or T), each with distinct cutoff points and percentiles. Troponin values are associated with clinical outcomes, but it is unclear whether either troponin offers better prediction of events. Purpose To evaluate the performance of Troponin I (TnI) and Troponin T (TnT) in ACS, to predict mortality, ischemic (Myocardial infarction, in stent thrombosis, stroke/TIA or post MI Angina) and bleeding events (BARC 2 or higher). Methods We analyzed a prospectively collected Argentinian multicenter registry called ReSCAR, which included patients with ACS. All adverse in-hospital events alongside occurrences within a one-year follow-up period were documented. We assessed the effectiveness of the first TnI and TnT value to predict in-hospital ischemic and bleeding events employing the area under the ROC curve. Additionally, we evaluated each troponin’s ability to predict total mortality within the one-year follow-up period. We also compared these performances with the GRACE and CRUSADE score. Results 983 patients were included. 291 had a TnI test (29.6%), 626 a TnT test (63.6%), while 66 values were lost (6.8%). 75.3% were male, 68.1% hypertensive, 25.9% diabetic, and 25.4% had a previous acute coronary syndrome. Mean age was 64.8 (95% CI 64-65.7), ejection fraction was 51.1 (95% CI 50.2-52), GRACE score 131.2 (95% CI 128.8-133.6), CRUSADE score 24.7 (95% CI 23.7-24.7). 38 patients died during hospitalization (3.9%), and a total of 71 died at 1 year follow up (7.2%). 81 patients (8.2%) had an ischemic complication at index hospitalization, and 29 (2.9%) a bleeding event. Median TnI was 85 (IQR 15-1209) while median TnT was 91 (IQR 21-635). TnT performance for predicting in hospital ischemic events, using area below the ROC curve was 0.63 (0.57-0.71) and TnI’s performance was 0.64 (0.53-0.76), p=ns, while GRACE score had a value of 0.67 (0.61-0.75) . For bleeding events, area was 0.71 (0.53-0.89) for TnI and 0.6 (0.47-0.73) for TnT, p=ns, compared to 0.67 (0.57-0.78) with the CRUSADE score . For in hospital mortality, TnI had a ROC curve of 0.54 (0.39-0.69) while TnTs was 0.63 (0.5-0.75), p=0.03, ), while GRACE's was 0.77 (0.68-0.86) (p=0.005). For total mortality at follow up, TnI had a ROC curve of 0.6 (0.5-0.72) compared to 0.63 (0.55-0.71) with TnT, p=ns, compared to 0.79 (0.73-0.85) with GRACE. Conclusion In our analysis, TnI and TnT had a similar, but discreet, predictive capability of in hospital ischemic events, that was similar to GRACE. Performance was slightly better, interestingly, for bleeding events, with TnI being numerically, but not statistically, better, while both were similar to CRUSADE. TnT outperformed TnI for the prediction of in hospital mortality, though this difference was lost when evaluating total mortality at follow up, and GRACE was superior to both.
Kersten et al. (Sat,) reported a other. Troponin I and T similarly predicted in-hospital ischemic events (AUC ~0.63), TnT better predicted in-hospital mortality (AUC 0.63 vs 0.54, p=0.03), GRACE was superior overall.